5.7 The Project Charter and the Problem Statement
Key Takeaways
- A project charter is the sponsor-to-team agreement authorizing the project, and it carries six blocks: business case, problem statement, scope, goal statement, performance measures, and team, resources, and milestones.
- A compliant problem statement answers six questions -- what, where, when, magnitude, baseline, and trend -- and states no cause and no solution, because naming an x pre-empts the Analyze phase.
- Baseline and trend are the two elements most often omitted: baseline is what the goal and the benefit are calculated against, and a deteriorating trend signals a special cause needing containment rather than a five-month DMAIC project.
- The ASQ Body of Knowledge spells SMART as specific, measurable, actionable, relevant, and time-bound, and it places the target and date in the goal statement, never in the problem statement.
- Charters name primary, secondary, and consequential metrics; the consequential 'do no harm' measure is what stops a team improving yield by slowing the line.
The project charter is the document that authorizes a Six Sigma project. It is the agreement between the sponsor, who commits money and people, and the Black Belt, who commits to a defined result by a defined date. Nothing in DMAIC is binding until it is chartered, and everything the team is later measured against is written here.
The Body of Knowledge places the problem statement at cognitive level Evaluate (IV.B.2) -- the highest level in the Define section. The exam will not simply ask what a problem statement contains; it will show you one and ask you to judge it against baseline performance and improvement goals. Practise reading them critically.
The charter's six blocks
| Block | The question it answers | Where this guide develops it |
|---|---|---|
| Business case | Why is this worth a Black Belt's capacity, and what happens if we do nothing? | Section 5.6 |
| Problem statement | What is wrong, where, since when, how big, and against what baseline? | This section |
| Project scope | Which process, sites, products, and period are in and out? | Section 5.8 |
| Goal statement | What target, by what date, measured how? | Section 5.9 |
| Performance measures | Which primary, secondary, and consequential metrics track the result? | This section |
| Team, resources, milestones | Who does the work, for how much of their time, against which toll gates? | Sections 5.10 and 5.11 |
A charter that is missing any one of the six is not a short charter; it is an unapproved project. The most commonly omitted block is performance measures, and the most commonly weak one is the problem statement.
The six elements of a problem statement
A compliant problem statement answers six questions. Four of them describe the defect; two of them anchor it to data.
| Element | Question | Example content |
|---|---|---|
| What | Which output ($Y$) is failing, in what unit? | First-pass yield on moulded housings |
| Where | Which process step, line, site, or segment? | Plant 2, moulding lines 3 and 4 |
| When | Over which period was this observed? | 26 weeks ending 30 June 2026 |
| Magnitude | How large is the gap, in the metric's own unit? | 91.4% against a 97.0% requirement |
| Baseline | What is the measured current performance, on what sample? | 91.4%, $n = 1.18$ million parts |
| Trend | Is it stable, deteriorating, or improving? | Stable, no trend across the 26 weeks |
Baseline and trend are the two that candidates leave out, and they are the two the exam tests hardest. Baseline is what the goal is subtracted from and what the benefit is calculated against; without it there is no way to prove improvement at closure. Trend changes the project entirely: a stable 91.4% is a capability problem calling for DMAIC, while a yield that fell from 97% to 91% in six weeks is a special cause calling for containment and a much faster investigation, not a five-month project.
The translation chain
Six Sigma solves a business problem by converting it into a statistical one and converting the answer back. The problem statement is the first link.
- Practical problem -- "Housings are being scrapped and we are missing shipments."
- Statistical problem -- "Mean wall thickness on lines 3 and 4 differs from nominal, and $\sigma$ exceeds the level required for $C_{pk} \ge 1.33$."
- Statistical solution -- "Barrel temperature and hold pressure explain 78% of the thickness variation; the optimum is 214 °C and 640 bar."
- Practical solution -- "Re-set and control the two parameters, verify capability, hand over."
This is the operational meaning of $Y = f(x)$: the problem statement names the $Y$ and refuses to speculate about the $x$'s. A statement that names an $x$ has skipped steps 2 and 3 and has decided the answer before collecting the data.
Building one, in three drafts
Draft 1 (unusable). "Quality on the moulding lines is poor and customers are complaining." No metric, no location, no period, no size. Nothing here can be measured or closed.
Draft 2 (better, still non-compliant). "First-pass yield on moulding lines 3 and 4 is too low because the second-shift operators are not following the setup procedure." It now has a $Y$ and a location, but it asserts a cause. If the setup procedure turns out to be irrelevant, the team has spent its credibility defending a sentence in the charter.
Draft 3 (compliant). "Over the 26 weeks ending 30 June 2026, first-pass yield on Plant 2 moulding lines 3 and 4 averaged 91.4% ($n = 1.18$ million parts) against the operating plan requirement of 97.0%. The gap is stable with no trend, and the associated scrap and rework cost is $162,120 annualized."
Draft 3 states what, where, when, magnitude, baseline, trend, and financial consequence, and it names no cause. Note also what it does not contain: a target and a date. Those belong to the goal statement (section 5.9), and merging them is a common charter defect -- the problem statement describes the present, the goal statement describes the commitment.
Evaluating a problem statement
Because this topic is tested at the Evaluate level, work from a defect list rather than a template.
| Defect | How it shows up | Consequence |
|---|---|---|
| Solution or cause embedded | "because…", "due to…", "we need a new gauge" | The Analyze phase is pre-empted; the real cause is never found |
| No baseline | An adjective instead of a number | Improvement cannot be demonstrated at closure |
| No period | "recently", "for some time" | The data set is undefined; anyone can dispute the sample |
| Unmeasurable $Y$ | "poor communication", "low morale" | No operational definition, no data, no project |
| Scope of the whole business | "order-to-cash performance" | A programme, not a project; it will never close |
| Goal folded into the problem | "reduce defects to 1%" | The baseline is lost and the target is unjustified |
| Trend ignored | Stable and deteriorating treated alike | DMAIC applied to a special cause that needed containment today |
| Cost stated with no basis | "costs us a fortune" | Finance disputes the benefit at closure |
A statement that survives all eight is defensible in a toll-gate review.
Performance measures in the charter
The charter names three kinds of metric, and the third is the one candidates forget.
| Metric type | Purpose | Example on this project |
|---|---|---|
| Primary | The $Y$ the goal is written against; the single measure of success | First-pass yield, % |
| Secondary | Supporting measures that explain movement in the primary | Scrap cost per thousand parts, rework hours |
| Consequential | The "do no harm" measures that must not degrade while the primary improves | Cycle time, overtime hours, customer defect escapes |
Without a consequential metric, a team can improve yield by slowing the line and inspecting more, then report a success that the plant manager experiences as a loss. Every charter should name at least one consequential measure with a limit written next to it.
Resource commitment and sign-off
The final charter block converts named people into committed hours, which is what makes the plan real.
| Role | Charter commitment |
|---|---|
| Sponsor / champion | Approves the charter, funds it, removes organizational barriers, chairs toll gates |
| Process owner | Accepts the improved process at handover; must be named at charter time, not at Control |
| Black Belt | Leads DMAIC execution, typically at 80-100% of available time |
| Green Belts / team members | Named individuals with a stated percentage of time, agreed with their line managers |
| Subject matter experts | Named, with the phases in which they are needed |
| Finance representative | Validates the baseline cost and, later, the realized benefit |
Two disciplines make sign-off meaningful. Time commitments are agreed with each member's line manager before the charter is signed, not announced afterwards -- an uncommitted team is the most common cause of a project that stalls in Measure. And the charter is a living document: baselines move when Measure-phase data arrives, scope narrows when stratification localizes the problem, and benefits change as solutions become concrete. Each revision is re-approved by the sponsor at a toll gate rather than edited quietly, which is the governance mechanism developed in section 5.10. Cost of poor quality categories and the hard-versus-soft savings distinction used in the business case block are covered in section 3.6.
Which element MUST be strictly excluded when authoring a Six Sigma problem statement?
A Six Sigma team is reviewing a proposed goal statement: 'Improve customer service response times as soon as possible.' Which SMART criteria are violated by this statement?
During a DMAIC project, team members suggest expanding data collection to include international call centers, which were not listed in the initial charter. What immediate action should the Black Belt take?